复杂环境中稳健的红外目标跟踪在自主导航、无人机探测、预警等方面具有重要研究意义。就经典粒子滤波红外目标跟踪算法中单一的灰度特征缺乏鲁棒性引起跟踪失效的问题，提出了一种基于特征融合的粒子滤波红外目标跟踪算法。结果表明，该算法能够从跟踪鲁棒性、准确性和实时性 3个方面实现稳健的红外目标跟踪。

Steady target tracking in complex environment is applied widely in guidance, unmanned aerial vehicles detection, and warning, etc. To solve the single gray robustness failure in infrared target tracking, particle filter tracking algorithm based on feature fusion is proposed, and the result shows particle filter tracking algorithm based on feature fusion can handle tracking in complex scene well in robustness, accuracy and real-time performance.